SEARCH

SEARCH BY CITATION

Abstract

  1. Top of page
  2. Abstract
  3. Methods
  4. Results
  5. Discussion
  6. Conclusions
  7. Acknowledgements
  8. References

Since the introduction of anti-Rhesus (Rh) D prophylaxis for RhD-negative women, other Rh and non-Rh red cell alloantibodies have become relatively more important and are now responsible for the greater proportion of haemolytic disease of the newborn. Anti-C and anti-E are the most commonly implicated non-D Rh antibodies in the pathogenesis of haemolytic disease of the newborn1. In 1977 Pepperell et al.2 reported the outcome of 44 women with anti-E. This is the only published series that investigates the implications of anti-E during pregnancy. The present report presents a retrospective study of the outcome of 122 pregnancies in which anti-E was the sole alloantibody detected.


Methods

  1. Top of page
  2. Abstract
  3. Methods
  4. Results
  5. Discussion
  6. Conclusions
  7. Acknowledgements
  8. References

The serological findings of all pregnancies managed between 1965 and 1994 in Newcastle upon Tyne, a regional referral centre for haemolytic disease of the newborn (HDN) were reviewed. Data on titres, amniotic fluid investigations, cord blood results, treatment and outcome were collected contemporaneously and recorded on separate record cards filed within the department of haematology at the Royal Victoria Infirmary as part of a wider study of HDN instituted in the early 1960s. Some missing pieces of data were obtained from the Regional Transfusion Centre (for titres) or from individual baby casenotes (for outcome). All cases in which anti-E was the sole antibody detected were included in the study. The maximum documented titre was recorded. Anti-E was detected and titres measured by the indirect anti-human globulin test using serial 1:2 tube dilutions of maternal serum reacted with saline suspensions of panel cells throughout the entire period of the study. In keeping with other rare, non-Rh antibodies, indications for examination of amniotic fluid were based upon algorithms derived from the study of anti-D HDN3. Spectrophotometric analysis of amniotic fluid was carried out as previously described4 and the results expressed as the ratio of percentage transmission at 520 nm: percentage transmission at 490 nm. At delivery, direct anti-human globulin testing of cord blood was performed in all cases, using saline suspensions of red cells in a tube test until the early 1990s. During the last three years of the study a commercial, micro-tube card technique was used. A positive result identified infants with HDN. Cord blood haemoglobin (Hb) measurements, where available, were also recorded. Severity of HDN was defined according to criteria developed for HDN caused by anti-D3: 1. Mild cases were positive for anti-human globulin and required no transfusion; 2. moderate cases were anti-human globulin-positive with cord blood Hb > 11 g/dL and required transfusion; 3. severe cases were anti-human globulin-positive with cord blood Hb < 11 g/dL or bilirubin > 85 mol/L and required transfusion; 4. very severe cases were anti-human globulin-positive with cord blood Hb < 7.5 g/dL, or hydropic. Statistical analysis was performed using the Pearson correlation coefficient (Prism 2.10 for Windows, GraphPad Software Inc).

Results

  1. Top of page
  2. Abstract
  3. Methods
  4. Results
  5. Discussion
  6. Conclusions
  7. Acknowledgements
  8. References

Over the 29 years of the study period, anti-E was identified as the sole maternal alloantibody in 122 pregnancies in 118 women. Sixty-two infants (51%) were born with a positive direct anti-human globulin test to 59 mothers. There was one stillbirth, unrelated to HDN, at 36 weeks of gestation with the maximum anti-E titre recorded as 1/4. There were no cases of hydrops fetalis. Of the 62 infants with a positive direct anti-human globulin test, HDN was classified as mild in 48 cases, moderate in eight, severe in five and very severe in one. Treatment was given to 20 of the affected cases (32%); 13 infants (21%) received an exchange transfusion, two of whom received further top-up transfusions; one infant received a single top-up transfusion, and the remaining six were managed with phototherapy alone. There was one neonatal death of an infant treated with exchange transfusion who died nine days later from congenital heart disease (single ventricle).

Figure 1 shows the relation between maximum recorded anti-E titre and clinical severity of disease in those 70 cases with complete data. The maximum anti-E titre was recorded in 43 of the 60 pregnancies which gave rise to unaffected (negative direct anti-human globulin test) infants. Titres ranged from 1/1 to 1/128 including four cases in whom anti-E was only detected by enzymatic techniques. These four have been assigned a maximum titre of 1/1. Maximum anti-E titres were recorded in 27 of the 62 pregnancies which resulted in affected (anti-human globulin-positive) infants. Titres ranged from 1/1 to 1/256 in the mild HDN cases, from 1/1 to 1/128 in the moderate HDN cases, from 1/1 to 1/1600 in the severe HDN cases and was 1/1 in the single very severe case.

image

Figure 1. Relationship between maximum anti-E titre and severity of haemolytic disease of the newborn (HDN) in 70 of the 122 pregnancies where anti-E was the sole alloantibody. HDN was classified according to Walker3.

Download figure to PowerPoint

In those who were transfused the maternal anti-E titre ranged from 1/1 to 1/1600, compared with 1/1 to 1/256 in those who were not transfused. Cord blood Hb levels were available from 26 anti-human globulin-positive cases (data not shown) There was no correlation between cord blood Hb and anti-E titres in these cases (Pearson correlation coefficient r=−0.26).

Amniotic fluid was analysed in seven affected pregnancies, with repeat testing in five. One case of severe disease (labour induced at 35 weeks of gestation) and three with mild disease were correctly predicted. Amniocentesis was falsely reassuring in one severely and one moderately affected case; liquor ratio fell from 1:081 at 30 weeks to 1:078 at 32 weeks, and from 1:067 at 30 weeks to 1:043 at 32 weeks, respectively (all values are within or at the upper limit of the mild/unaffected zone4).

Discussion

  1. Top of page
  2. Abstract
  3. Methods
  4. Results
  5. Discussion
  6. Conclusions
  7. Acknowledgements
  8. References

Although this study was initially conceived as part of a prospective study of HDN, this collation of data constitutes a retrospective study. Its major failing is the lack of data on anti-E titres in 52 of the 122 pregnancies. Nevertheless, there are sufficient data to make analysis worthwhile, and it remains the largest to assess the fetal risks associated with maternal anti-E alloanti-bodies. Considerable overlap in maximum anti-E titres between anti-human globulin-positive and negative infants, combined with lack of correlation between anti-E titre and both cord Hb and HDN severity, restricts the predictive power of anti-E measurements. Due to the paucity of data on the rare anti-red cell antibodies, it is common practice for units either to follow guidelines derived from data on RhD alloimmunised pregnancies or, conversely, to dismiss the presence of anti-E in particular as being of little consequence. The temptation to dismiss anti-E is increased because it may be a naturally occurring antibody.

In this study, because of the lack of any substantial data at the time it was set up, no decisions on the mode or timing of delivery were based on anti-E titres. In only one case was labour induced prematurely based on the results of amniotic fluid examination. This infant had severe disease. No fetal blood samples were taken and no fetus received intrauterine transfusion. The use of a potentially more sensitive anti-human globulin test in the last three years of the study will have had a negligible effect on the total number of affected infants. Future use of more sensitive techniques over another 30 years might increase the total number of affected infants detected and thus might reduce the proportion (but not absolute numbers) of such infants requiring treatment.

The data from this study, which is effectively a documentation of the natural history of anti-E HDN in pregnancy, show that a substantial proportion of infants are sufficiently affected by anti-E that they suffer from clinically important HDN. Twenty-one percent of affected infants received exchange transfusion and 10% had severe or very severe disease. The single, very severely affected case had a maximum anti-E titre of 1/1, highlighting the disparity between anti-E titre and disease severity. A co-existent, non-immune cause for anaemia, however, could not be definitively excluded in this case.

When paternity is certain, the probable genotype of the father will provide some guidance about the risk of HDN. The maternal anti-E titre, however, correlates poorly with observed severity of disease in her infant. In this respect, the position resembles HDN due to anti-Kell5, but not that due to anti-D. However, serious morbidity may well be more common in both these two diseases. The limited data available suggest that spec-trophotometric examination of amniotic fluid is also a poor indicator of outcome. Its role in management of these pregnancies is therefore questionable. In particular there are no data in this study on hydropic infants or stillbirth, the two truly important consequences of HDN which might impel us to intervene in a pregnancy. Thus prediction of pregnancy outcome remains difficult. Amniocentesis with polymerase chain reaction can now be used to determine RhE/e genotypes6 and may have a role in the future management of this disease. Such tests might provide rather more useful data than spectrophotometric analysis of amniotic fluid. It remains unknown whether the presence of anti-E exerts a cumulative effect when combined with other antibodies. Fetal blood sampling is able to provide a definitive fetal genotype, confirmation of the diagnosis of HDN and an acceptable measurement of severity. However the indications for fetal blood sampling in HDN due to anti-E are unclear. The data in this retrospective study do not allow us to recommend a specific management policy; in particular it is not possible for us to formulate an evidence-based policy for the use of fetal blood sampling in this disease.

Conclusions

  1. Top of page
  2. Abstract
  3. Methods
  4. Results
  5. Discussion
  6. Conclusions
  7. Acknowledgements
  8. References

In summary, the two main conclusions of this study are that anti-E titres are an insensitive and poor predictor of subsequent HDN severity, and that it would be unwise routinely to dismiss anti-E as being of little clinical consequence. We can find no grounds for assuming that protocols developed for the management of HDN due to anti-D are appropriate for women with anti-E. We do recommend that all infants born to mothers with anti-E of whatever titre should be considered potentially at risk of HDN, and that a direct anti-human globulin test and haemoglobin measurement should be performed on cord blood, unless the fetus is already known to be E negative.

Acknowledgements

  1. Top of page
  2. Abstract
  3. Methods
  4. Results
  5. Discussion
  6. Conclusions
  7. Acknowledgements
  8. References

We would like to acknowledge the late Professor W. Walker whose prospectively collected data provided the opportunity for this study, and thank Ms J. Gibson who extracted most of the serological data from his records.

References

  1. Top of page
  2. Abstract
  3. Methods
  4. Results
  5. Discussion
  6. Conclusions
  7. Acknowledgements
  8. References